Maybe We Should Let Machines Pick Our Leaders

In light of recent world events it may be time to consider that letting people run countries based on their ability to get elected may not be the best way to do things. Perhaps it’s time to let computers decide who is best equipped to be in charge of these enormous economies and bureaucracies. Handing the choice over to an unthinking machine algorithm may seem like a crazy alternative, but would it be any worse than the way we do things now?

Humans have experimented with many different ways to select leaders, including elections, heredity, and military coups. But nobody has ever trusted the running of a country to a machine before. Except, of course, in fiction.

This idea goes back at least to The Machine Stops (1909) by E.M.Forster, wherein humans can no longer live on the surface of the planet. They live below ground, their needs met by the Machine, an omnipotent computer.

More recently, the mind-bending novel If Then (2015) by Matthew De Abaitu features a small town in England where the lives of its citizens are controlled by a set of algorithms known as the Process. The Process is intended to balance the needs of all the citizens and give their lives meaning. Following a societal collapse brought on by the rise of algorithms, the characters in the novel decide to put their trust in these algorithms, handing over their free will and agency when they do.

The Process is depicted as a higher power, one able to take a wider, more accurate view of what people in society actually need. This is at times in opposition to what they think they want, and comes at a price; some citizens are violently evicted from the village for reasons known only to the Process.

These fictional societies are often depicted as dystopian, the citizens within them having forfeited their self-determination. But do humans have some quality that makes their opinions superior to the analysis of machines? And what if belief in human opinion is the problem? That it blinds us to what is actually needed to effectively run a country? Take human bias out of it and we might wind up with some people that are actually qualified for the job.

Cynically, it can appear that the most important ability of today’s politicians is getting elected. In what other field of human endeavour does having the skills to do the job count for so little? You don’t pick a heart surgeon based on their winning smile. Politicians are doing a job, after all: guiding the writing of laws and directing the spending of taxes to benefit the state and its citizens. Is popularity the most important requirement for this job?

A computer algorithm could be taught to identify people who have the actual skills needed to run a country from digital profiles. They could sift through literally millions of potential leaders and winnow it down to a list for citizens to pick from. All for a fraction of the cost and effort it currently takes to pick candidates.

A common theme of The Machine Stops and If Then seems to be that if humans hand the running of societies over to machines, then the meaning of their lives is somehow diminished. But algorithms already make so many choices for us, and we don’t appear to have a problem with that.

We’re quite happy to let an algorithm guide our choices of what music we listen to, what videos we watch, and what consumer products we buy. We’ll let an algorithm pick stocks for us or even design a concert hall. Why not let one shortlist viable and competent political leaders? As Pedro Domingos points out in The Master Algorithm (2015), Barack Obama was able to gain an edge over Mitt Romney in the 2012 US election by using machine learning algorithms. Political parties will increasingly use computers to analyse the electorate and direct their resources.

Maybe we’ve got that the wrong way around; maybe computers should be used to benefit the populace, rather than the politicians.